The Plan S criteria have definitively softened over time away from the original aim to force a hard OA policy for research funded by the signatories, and to break the power of the big commercial publishers to charge handsomely, either for subscriptions behind paywalls, or as APCs . They have really backed away from limiting the high APCs charged by commercial publishers referred to @15min [the rest of the sector charges lower APCs or nothing, with exception of a few societies in STEM]. They say “the journal/platform must provide APC waivers for authors from low-income economies and discounts for authors from lower middle-income economies, as well as waivers and discounts for other authors with demonstrable needs. Waiver policies must be described clearly on the journal website/platform and statistics on waivers requested and granted must be provided.” and there are other statements about the requirement from publishers for transparency about the APCs charged, and that funders should be paying the APCs [with some wiggle room]. https://www.coalition-s.org/addendum-to-the-coalition-s-guidance-on-the-implementation-of-plan-s/principles-and-implementation/

In the academic-led no budget journal sector [yep we are interested in ‘changing the world’ @ 29min], we get very little from Plan S. No money for xml machine readability, or clockss style archiving. We asked. Some help with ‘transformative agreements’ perhaps, which as you note at 19min, advantages the big players more as currently structured. See https://www.coalition-s.org/guidance-on-the-implementation-of-plan-s/

So, 2 cheers for Plan S: data and articles are at least becoming more available to the public. Publishers have still been able to maintain those profits, and placate shareholders.